Originally Posted by
jsloan
Well, if I understand the initial conclusions properly, it sounds like what happened is that the computer was programmed for a course of 280 degrees based upon the expected departure on 28R. This should have been corrected to a desired course of 010 degrees, based upon the actual departure on 1R, but it apparently was not. Thus, when the computer was engaged, it tried to come to 280 degrees, which would involve a left turn from a 010 heading. So, I think it was less “the pilots deliberately turned left” and more “the pilots activated the autopilot and the autopilot turned toward the programmed heading.” So, if there is a systemic failure here, I imagine it will be something on the checklist regarding verifying the computer has the correct departure runway programmed in.
Yes, I fully understand that. What I don't understand is why one of them didn't immediately realize something was wrong. You just don't turn left immediately after a 1R takeoff. The entire West departure plan is premised on parallel takeoffs spaced in between landings on the 28s. Turning left from 1R is akin to turning left from the right lane in heavy traffic - you are almost certainly turning directly into the path of another vehicle (or in this case, another aircraft).